14:03:04 <SridarK> #startmeeting fwaas 14:03:05 <openstack> Meeting started Tue Sep 5 14:03:04 2017 UTC and is due to finish in 60 minutes. The chair is SridarK. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:03:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:03:10 <openstack> The meeting name has been set to 'fwaas' 14:03:16 <reedip_> SridarK : either you forgot how to start a meeting or somethings down 14:03:23 <reedip_> oh you forgot how to start the meeting :P 14:03:35 <SridarK> SridarK: sorry typo early AM :-) 14:03:41 <SridarK> i need to wake up 14:03:50 <reedip_> Coffee !!! 14:03:56 <SridarK> yes badly needed 14:04:12 <SridarK> #chair xgerman_ yushiro 14:04:13 <openstack> Warning: Nick not in channel: yushiro 14:04:14 <openstack> Current chairs: SridarK xgerman_ yushiro 14:04:22 <doude> o/ 14:04:28 <SridarK> i forget who was to run the mtg today 14:04:39 <reedip_> you are ... I think 14:04:50 <SridarK> ok lets get started 14:05:13 <SridarK> #topic Queens 14:05:20 <yushiro> Hi 14:05:28 <SridarK> #chair yushiro xgerman_ 14:05:29 <openstack> Current chairs: SridarK xgerman_ yushiro 14:05:30 <yushiro> I'm sorry I'm late. 14:05:36 <SridarK> yushiro: np 14:05:58 <SridarK> so in terms of Queens - most important is the planning and PTG topics 14:06:28 <SridarK> #link https://etherpad.openstack.org/p/neutron-queens-ptg 14:06:59 <SridarK> pls add to the existing list in the fwaas section 14:07:37 <SridarK> shall we take a few mins to discuss on any other priorities as well now ? 14:08:08 <reedip_> SridarK : dont we have an etherpad for FWaaS only ? 14:08:13 <reedip_> I think xgerman_ made something 14:08:26 <SridarK> I am open to have a targetted mtg on Thu as well either in our channel or as a conf call or something too 14:08:32 <xgerman_> I think I did 14:08:59 <xgerman_> https://etherpad.openstack.org/p/fwass_ptg_denver 14:09:09 <reedip_> SridarK : meeting/conf call would be great 14:09:26 <reedip_> #link https://etherpad.openstack.org/p/fwass_ptg_denver 14:09:54 <SridarK> reedip_: ok - shall i set something up for the same time in 48 hrs ? 14:10:15 <reedip_> Oh , I thought during the PTG :P 14:10:32 <reedip_> a normal meeting might suffice before the PTG 14:10:58 <xgerman_> PTG might be hectic… 14:11:07 <SridarK> reedip_: ok - i can do that for sure - but my sense is the audio quality is going to be bad 14:12:05 <SridarK> lets take a quick stab now 14:12:14 <reedip_> xgerman_ SridarK : I think on the last day when we have the time only for sub teams, we can have the discussion , yushiro would be there with you, so if not audio, we can just jump in the #fwaas channel and the FWaaS etherpad 14:12:45 <SridarK> of the priorities already listed in #link https://etherpad.openstack.org/p/neutron-queens-ptg 14:12:55 <SridarK> what do folks feel ? 14:13:02 <SridarK> reedip_: +1 we can do that 14:13:38 <reedip_> folks are pretty quiet .... :P 14:13:56 <SridarK> I think in terms of debt or things needing wrap up - we have clarity 14:13:59 <xgerman_> after a US holiday… 14:14:05 <annp> +1 reedip_ 14:14:10 <yushiro> SridarK, If possible, would you mind adding 'logging extension' as a challengable topic? 14:14:12 <SridarK> L2 support, Horizon, testing 14:14:31 <SridarK> yushiro: yes sure 14:14:38 <xgerman_> + 1 14:15:02 <SridarK> yushiro: done 14:15:09 <yushiro> SridarK, thank you so much. 14:15:16 <annp> Thanks SridarK 14:15:16 <SridarK> yushiro: np 14:15:25 <SridarK> annp: np 14:15:31 <yushiro> SridarK, xgerman_ I agree with current topic. L2 is the highest priority :) 14:15:50 <yushiro> SridarK, xgerman_ and horizon ;) 14:15:53 <SridarK> i think we also need to evaluate ovs and sg integration 14:16:01 <xgerman_> +1 14:16:06 <chandanc> +1 14:16:26 <yushiro> SridarK, ah, yes OVS one is also necessary, isn't it? chandanc :) 14:16:37 <xgerman_> also investigate what happened to FW next and how we can collaborate 14:16:47 <xgerman_> SG 14:16:47 <chandanc> ya, i agree 14:17:36 <SridarK> chandanc: , yushiro xgerman_ and myself can discuss with the right set of folk on ovs 14:17:45 <xgerman_> +1 14:17:53 <chandanc> sure 14:17:57 <reedip_> I rearranged the things a bit in the etherpad as per priority 14:18:07 <yushiro> SridarK, +1 jakub and other OVS guy can attend PTG 14:18:25 <SridarK> chandanc: pls let us know if there are specific concerns that u want clairfication on as well 14:18:41 <chandanc> Sure, i had a look at the code and UT 14:18:58 <chandanc> i see a lot of changes has been done for the SG driver 14:19:32 <SridarK> chandanc: ok - can u pls take a look and send us a summary by the end of the week or early next week 14:19:36 <chandanc> I will try to resync 14:19:44 <xgerman_> thanks 14:19:44 <SridarK> chandanc: ok perfect thx 14:19:47 <chandanc> Ok SridarK 14:20:32 <SridarK> and then we have testing 14:20:44 <SridarK> this will keep us busy for a good part of Queens 14:20:56 <xgerman_> +1 14:21:24 <SridarK> In terms of new features, SFC integration & using CCF 14:21:35 <xgerman_> yep 14:21:52 <xgerman_> and then our long term plan hitiching our wagon to K8 14:22:33 <SridarK> I also reached out to CCF folks and did some investigation - i think they will have some model for a PoC - that we can play around with 14:22:40 <SridarK> xgerman_: huge +1 14:22:49 <reedip_> I am talking to them on #openstack-meeting :P 14:22:55 <reedip_> SridarK ^^ 14:22:56 <yushiro> +1 14:22:56 <xgerman_> :-) 14:23:14 <xgerman_> Yeah, integrating them should be stroght forward 14:23:18 <SridarK> i dont know how we position ourselves on that tha front 14:23:36 <SridarK> should we look at Kuryr as a first step too 14:23:41 <reedip_> SridarK: SFC would be something to work with... 14:23:51 <apuimedo> SridarK: you summoned the kuryr man 14:24:00 <xgerman_> nice 14:24:00 <reedip_> apuimedo : hey hi :) 14:24:04 <SridarK> I dont have much thoughts on that yet - but lets talk more 14:24:05 <xgerman_> hi 14:24:11 <SridarK> apuimedo: hi 14:24:26 <apuimedo> SridarK: may I help? 14:24:31 <SridarK> apuimedo: wow like a Genie - u surface :-) 14:24:39 <xgerman_> we were wondering how FWaaS can be useful for Kuryr 14:24:52 <SridarK> apuimedo: ^^^ exactly as xgerman_ puts it 14:24:59 <reedip_> IIRC Kuryr doesnt have the FWaaS extension yet, right ? 14:25:09 <apuimedo> xgerman_: well, it may be that for some policy in kuryr-kubernetes network policy translation fwaas would be needed 14:25:22 <apuimedo> we still didn't check if SGs are enough of not 14:25:37 <xgerman_> yes, that’s my thought as well. Canal, etc. are pretty poor in the semantics 14:25:37 <SridarK> apuimedo: perhaps we can thrash out some workflow and usage models 14:25:41 <apuimedo> it is a queens goal 14:25:50 <SridarK> apuimedo: will u be at the PTG ? 14:25:51 <xgerman_> oh, stars aling… 14:26:05 <apuimedo> SridarK: no. Denver + Sydney would be too much 14:26:14 <apuimedo> but dmellado is going on behalf of Kuryr 14:26:18 <apuimedo> he'll be the kuryr man then 14:26:28 <dmellado> o/ 14:26:31 <xgerman_> ok, so we need to go to Sydney to meet you? 14:26:36 <reedip_> lol 14:26:38 <SridarK> apuimedo: ok - we will try to sync up then 14:26:41 <SridarK> :-) 14:27:01 <dmellado> apuimedo: if you try to throw more stuff into me I'll exchange my name tag at the PTG with somebody xD 14:27:04 <yushiro> I want to go Sydney either... 14:27:07 <SridarK> dmellado: perhaps we can figure out a way to meet up to discuss some FWaaS usecases 14:27:15 <dmellado> SridarK: sure thing 14:27:16 <reedip_> yushiro : you are in line ! 14:27:34 <SridarK> dmellado: yushiro, xgerman_ and myself will be at Denver 14:27:34 <reedip_> apuimedo : kuryr meeting is on Mondays, right ? 14:27:56 <dmellado> SridarK: I'll be at the infra sessions on Monday, so feel free to catch me there 14:28:11 <SridarK> dmellado: i get in only on Tue eve - 14:28:22 <SridarK> and there till Fri 14:28:23 <dmellado> SridarK: oh, I see, np I'll be there the whole week 14:28:29 <apuimedo> reedip_: it is 14:28:31 <xgerman_> me, too 14:28:39 <SridarK> dmellado: ok perfect - we can meet up on Wed 14:28:41 <dmellado> ping me and we'll try to meet up there at some point 14:28:54 <SridarK> dmellado: perfect done 14:30:13 <apuimedo> :-) 14:30:16 <SridarK> ok good we have a plan to explore 14:30:23 <SridarK> thx apuimedo dmellado 14:30:31 <dmellado> yw! 14:30:41 <xgerman_> thx 14:30:45 <yushiro> thank you 14:30:54 <reedip_> ea 14:31:30 <apuimedo> thank you 14:31:47 <SridarK> reedip_: yes on SFC, as u mentioned 14:32:13 <reedip_> sorry SridarK : lost track 14:32:28 <yushiro> apuimedo, dmellado looking forward to meeting in Denver :) 14:32:43 <SridarK> I think we have a reasonable list to go after, other pls add things that are important that we have missed 14:32:45 <reedip_> I am discussing for CCF. SFC not yet 14:33:14 <SridarK> reedip_: sounds good - i had some conversations on SFC in the past 14:33:30 <reedip_> ok .... 14:33:49 <dmellado> yushiro: same there! 14:34:18 <SridarK> ideally if we can wrap up our debt from the Pike cycle and move fwd on 1 or 2 of these new areas - will be ideal 14:34:21 <reedip_> SridarK : I do not think there would be some exceptional work for SFC. I think it would just be pluggable 14:35:01 <SridarK> reedip_: that was my thought too - i had some concerns on the implementation details 14:35:02 <xgerman_> the devil is in the details 14:35:12 <reedip_> I am checking the details :P 14:35:12 <SridarK> xgerman_: yes exactly 14:35:34 <SridarK> ok good i think we have enough to keep us busy during PTG 14:35:42 <SridarK> lets move on 14:35:45 <xgerman_> they run a port in-different port out model,,, 14:36:15 <SridarK> xgerman_: and based on how we are positioned in the router - i was not sure how to effect it 14:36:17 <reedip_> SridarK , xgerman_ have we considered providing support to tacker ? 14:36:18 <reedip_> for NFV ? 14:36:40 <reedip_> or is it already there ? 14:36:57 <SridarK> reedip_: we had some very early conversations but no discussions recently 14:37:06 <xgerman_> +1 14:37:08 <reedip_> xgerman_ https://docs.openstack.org/newton/networking-guide/config-sfc.html#architecture 14:37:25 <amotoki> "for NFV", "for tacker" sounds too ambiguous to me. 14:37:36 <reedip_> SridarK : ok, I think that work would be done from tacker team themselves 14:37:41 <amotoki> it looks better to discuss specific topics if there are such demands 14:37:49 <xgerman_> +1 14:37:52 <reedip_> amotoki +1 14:38:17 <SridarK> ok sounds good and agree 14:38:23 <SridarK> #topic L2 Support 14:38:35 <SridarK> yushiro: chandanc pls go ahead 14:38:41 <yushiro> SridarK, OK. 14:38:51 <yushiro> #link https://review.openstack.org/#/c/323971/ 14:39:45 <yushiro> I got some comment from Inessa and he said 'imho noop driver dependency has a lot more chances to get approved and merged.' 14:40:57 <reedip_> yushiro : he reason for that is also mentioned 14:41:02 <reedip_> and I think he is right 14:41:19 <reedip_> "Maybe make this patch independent of ovs driver by implementing noop driver? 14:41:20 <reedip_> ovs driver patch https://review.openstack.org/#/c/447251/ that this patch depends on is a very complex change, imho noop driver dependency has a lot more chances to get approved and merged." 14:41:32 <yushiro> Yes, I'd like to discuss with a behavior when 'noop' driver is loaded. 14:42:32 <yushiro> I think 'noop' driver does nothing and alarm some message(WARNING) into log, right? 14:43:30 <amotoki> yushiro: why do you think some warning log messages are needed? 14:43:57 <reedip_> No , it doesnt ... 14:44:00 <yushiro> amotoki, Because, if 'noop' driver is loaded, firewall rule doesn't set in OVS. 14:44:08 <amotoki> noop driver does nothing on back-end side and if the noop dirver is loaded just API should work as usual 14:44:18 <reedip_> https://github.com/openstack/neutron/blob/master/neutron/services/metering/drivers/noop/noop_driver.py 14:44:32 <SridarK> In the past, for v1 as well as v2 (L3) we have not taken such an approach - the agent - driver interface is basically the internal API corresponding to actions for CRUD 14:44:54 <yushiro> amotoki, reedip_ Aha. 14:45:19 <SridarK> so this will be a pass through 14:45:31 <reedip_> I guess so 14:45:41 <yushiro> amotoki, reedip_ SridarK OK, thank you and I was misunderstanding. 14:45:59 <amotoki> re: noop driver in the agent side, it is in a question how it has a real value though.. 14:46:14 <amotoki> but it can be used to validate internal APIs 14:46:47 <SridarK> amotoki: perhaps that and the removal of dependency with the driver patch - 14:46:55 <reedip_> amotoki : I think we can keep noop driver for now, and once OVS Firewall patch merges, then maybe we can add this back 14:47:08 <amotoki> SridarK: +1, yeah it is one possible usecase 14:47:30 <SridarK> yushiro: since our driver interface is well defined - this should not be a problem 14:47:53 <SridarK> perhaps it gives folks who want to use a different driver more confidence in integration 14:48:34 <yushiro> SridarK, amotoki OK. So, we should implement driver base code and define some methods, shouldn't we? 14:48:35 <SridarK> yushiro: anyways we allow specifying the driver 14:49:10 <yushiro> SridarK, I understood. 14:49:23 <amotoki> yushiro: the base driver is useful to deifne the interface 14:49:37 <amotoki> with docstring 14:49:57 <yushiro> So, I'll remove ovs = neutron_fwaas.services.firewall.drivers.linux.l2.openvswitch_firewall.firewall:OVSFirewallDriver in setup.cfg from l2-agent patch. 14:50:07 <yushiro> amotoki, Yes, I guess so! 14:50:31 <SridarK> yushiro: ok 14:50:46 <yushiro> chandanc, please add above definition into setup.cfg in your OVS patch :) 14:50:54 <SridarK> we have 10 mins, chandanc anything to add from driver side ? 14:51:06 <yushiro> chandanc, I'll comment in gerrit either. 14:51:10 <chandanc> yushiro: sure i agree 14:51:21 <SridarK> ok lets move on 14:51:30 <yushiro> yes 14:51:32 <SridarK> #topic Horizon changes 14:51:41 <SridarK> #link https://review.openstack.org/#/c/475840/ 14:51:52 <SridarK> SarathMekala: amotoki pls go ahead 14:52:14 <SarathMekala> I returned back from vacation today 14:52:21 <SridarK> SarathMekala: do u feel we are ready to do more manual tests 14:52:23 <SridarK> SarathMekala: ok 14:52:27 <yushiro> SarathMekala, Welcome :) 14:52:35 <SarathMekala> :) thanks 14:52:38 <amotoki> re v2 dashboard, SarathMekala post a patch with UT success :) yay 14:52:45 <SarathMekala> as of the last checkin the UTs are passing now 14:53:46 <amotoki> perhaps some more test round will be needed 14:53:57 <SridarK> amotoki: thx for jumping in on the UTs as well 14:54:00 <SarathMekala> there are some comments given by amotoki.. will work on them 14:54:35 <SridarK> SarathMekala: amotoki ok - i will pick up PS28 14:54:38 <SarathMekala> now we can do one big round of manual testing and fix any issues that turn up 14:54:53 <SarathMekala> this should harden the flows :) 14:55:06 <yushiro> #link https://etherpad.openstack.org/p/fwaas-v2-dashboard 14:55:41 <SridarK> ok we can update the etherpad ^^ 14:56:03 <SarathMekala> yes.. will clean it up a bit 14:56:04 <yushiro> SarathMekala, I found GUI issue in selecting firewall_rule in policy. Please refer #fwp-6 14:56:43 <SarathMekala> yushiro, ok will check it out 14:56:59 <amotoki> yushiro: SarathMekala: firewall rule selection in policy creation seems not to work at now 14:57:01 <SridarK> SarathMekala: ok pls evaluate that and give us a heads up 14:57:35 <amotoki> what I can tell is that it is due to django 1.11 stuff and v1/v2 dashboard have the same problem 14:58:00 <yushiro> amotoki, Oh, OK. Thanks. 14:58:03 <amotoki> it is caused by horizon side code, but in the horizon side, we are dropping the consumer of the corresponding JS code 14:58:33 <amotoki> so if the fw dashboard continues to use it, we need to maintain and/or fix the probem in the fwaas team 14:58:54 <SridarK> ok 2 mins, let me go to open slot and we can continue discussion 14:59:01 <yushiro> amotoki, you mean v1 dashboard? 14:59:01 <SridarK> #topic Open Discussion 14:59:10 <amotoki> yushiro: including v1 dashboard 14:59:17 <SridarK> others anything else to bring up ? 14:59:17 <yushiro> amotoki, Ok 14:59:37 <amotoki> some maintenance notice: regarding v1 dashboard I sent a series of patches which I noticed during v2 dashboard reviews. v2 dashboard copied a lot from v1 dashboard including some weird codes. https://review.openstack.org/#/q/status:open+project:openstack/neutron-fwaas-dashboard+branch:master 15:00:21 <yushiro> amotoki, wow, nice cleanup! will take a look. 15:00:25 <SridarK> ok we are nearing end - we can continue the fwaas channel 15:00:29 <doude> Hi I just started to rebase patch to implement a plugable backend driver 15:00:32 <SridarK> yushiro: +1 15:00:38 <doude> https://review.openstack.org/#/c/480265 15:00:39 <SridarK> amotoki: will look too 15:00:43 <SarathMekala> amotoki, +1 15:00:47 <SridarK> doude: ok 15:01:00 <SridarK> doude: will add to the agenda so u have time 15:01:10 <SridarK> #endmeeting fwaas